Build WiFi Range Extender from 18" Parabolic Dish, construction pictures + instructions

Build A Tin Can Waveguide WiFi Antenna In One Hour. Then put it on a
18" inch direct tv dish, (optional). No coax signal loss. I use a USB
device on antenna, so only a usb cable feeds my computer. With 16' USB
REPEARER CABLES end-to-end you can run 80 feet of cable without loss.
Complete instructions. This Powerful wlan antenna REALLY KICKS BUTT!!


Here is my version of the dish conversion to wifi range extender->
http://www.findmorecollectibles.com/wifi.html

enjoy,
ac6ts

 

Joe,
I am only running 32 feet, (2 cables), from my living room to my radio
room dish. I quoted the 80 ft specs from the packages, but have not
tried that run distance. No need here. I just took their word for it.
Thanks for your info. That is important.

john
http://www.findmorecollectibles.com/wifi.html

 

Nicely done. You might want to add some numbers.

Maximum gain for a 0.6meter diameter dish:
gain = 9.87 * Dia^2 / wavelength^2 * (feed efficiency)
gain = 9.87 * 600mm^2 / 125mm^2 * 0.4
gain = 91
dBi = 10 log(91) = 19.5dBi
The 40% efficiency (that's the 0.4) is probably optimistic.

However, that's the best that can be done. With a mismatched feed
(your USB contraption), you'll be radiating much of the signal into
all manner of undesireable directions, with very little hitting the
dish and being reflected in the desired direction. That's why feed
point "illumination" is important. Ideally with a properly matched
feed, you'll get the same gain in xmit and recv. However, with your
USB device feed system (an almost hemispherical feed pattern), all of
the signal arrives at the feed in receive, but much of the signal goes
nowhere useful in xmit. So, you'll get somewhat less than the
theoretical 19.5dBi gain in receive, but xmit will be far less gain.

Might was well guess how much less. Assuming a hemispherical pattern
from the USB antenna, only about 10% of the RF hits the dish. The
rest goes off in useless directions. So, 10% is:
loss = 10 log(0.1) = -10dB
So, my guess is that the antenna gain in xmit is about:
19.5 - 10 = 9.5dB
Yech. (I can produce a more accurate number for the 10% later but I
don't have a DBS dish to measure handy right now).

 

Sigh. Sorry, I goofed. I was looking at a variety of DBS dish
reflectors on other sites and got your construction mixed with another
one that did NOT use a can type feed. My comments about different
gain in xmit and receive apply only to building it without the tin
can. If the can feed illumination pattern is reasonably well matched
to the diameter of the dish, the tx and rx gains should be the same.
My appologies for the wrong info.

 

The tin can was cut to 3-1/2 inches, which seems too short to me.
with a 4" diameter can, I think the radiator should be at 1.7", but I would
like to see the can at 5.5" as a standalone can.

Should the radiator be at the focal point, or the mouth of the can
(properly sized for its own radiator)?

And where might we assume that the radiator is on that adapter? It looks
like it might be in the middle of the bore, which is good, but that's an
awfully large hole in the side of the can, which I thought Jeff didn't like
in building a wire cantenna.

Bob Alston's coffee can
http://members.cox.net/tulsaalstons/Computer.htm
Clarence Dold's coffee can
http://www.rahul.net/dold/clarence/usb-can/im000742-800x600.jpg

 

3.5 inches might be a good length. With dish feeds, the idea is to
design the feed illumination angle so that it exactly covers the dish
area. No more, no less. In this case, there's no interest in
maximizing the gain because the dish does most of that, not the feed.
However, if he were trying to squeeze every last dB of gain out of the
antenna, feed gain would be important.

See:
<http://www.w1ghz.org/antbook/contents.htm>
<http://www.w1ghz.org/antbook/chap6-3.pdf>
See 6.3.1 above for details on the spillover (overspray) problem
and the introduction at:
<http://www.w1ghz.org/antbook/chap6-0.pdf>
for general clues.

New stuff:

Just inside the mouth of the can is at the focal point. The exact
location is the "phase center" of the feed. See fig 6.1-1 in:
<http://www.w1ghz.org/antbook/chap6-1a.pdf>
I gave up trying to calculate the exact location and just move the
feed in and out until the gain is at maximum.

I have a DWL-120 (somewhere) which is what the author used. It looks
like a loop antenna, but it's not. It's just a 1/4 wave (3.13cm)
piece of wire inside the plastic loop. The problem is that I don't
remember which side of the loop had the antenna. If I get ambitious,
I'll tear it apart and see. Meanwhile, you're correct. The antenna
is *NOT* in the center of the loop and therefore in the canter of the
can. That will cause a lopsided illumination pattern.

My (paper) coffee can:
<http://802.11junk.com/jeffl/antennas/coffee2400/>
I could probably scale the model to match the tin can in the DBS dish
construction, and see how it looks. (yet another project).

My hacked version of Ivor Hewitt's Excel spreadsheet for generating an
NEC2 "card" suitable for feeding to 4NEC2 antenna modeling software:
<http://802.11junk.com/jeffl/rf-calc/Coffee-can05.xls>
Unfortunately, it has some problems that I've been too lazy (or too
inept) to fix. Caveat Emptor and let me know if you find any (more)
bugs.

 

My usb device started fruiting out and was very unstable after 2 repeaters.
I had to run a 2nd wire for the power.


Joe.

 

Arithmetic failure, not karma failure.

Let's play with the numbers....

You have a USB device that sucks perhaps 450ma at 5VDC at the end of
perhaps 25m of USB cable (5 lengths of 5m cable). Amplifiers take
care of data loss but not the DC power loss. Good USB cable uses
#20AWG stranded wire at 0.034 ohms/m. Cheap junk super flexible cable
uses #28AWG stranded at 0.232 ohms/m and isn't even worth calculating.
Also, each gold plated connection in line will also lose about 10mv.

So, the IR loss for #20 is:
2 wires * 25 m * 0.034 ohms/m * 0.45A = 0.77 VDC
Add to that the connector losses:
2 * 5 connectors * 0.01 v/connector = 0.10 V
for a total loss of 0.87 VDC. That means your USB radio would need to
run off about 4.1VDC, which probably won't work. I've played with
older USB radios and found that they'll sorta work at 4.5VDC, but tend
to do weird things. For example, since they draw more power in xmit
than in receive, it usually finds access points, but fails to connect.
The newer USB radios are actually 3.3VDC devices internally, and just
might work if the LDO (low dropout) regulator inside is doing its job.
That might be why the WG111V2 worked and the older model did not. Next
time I have an opportunity, I'll put a few USB radios on an adjustable
voltage power supply and see what happens.

 

A quick test on the ground showed all well, but after i installed everything
and ran cables thru the roof,down walls, ect.. i got the problems >.< my
WG111 seems to draw the max current for a port so it needed a little help
after the active cables chewed off a little current. The WG111v2 seems
stable without the help but its useless for long distance links. The extra
wire was only thin but did the job.

Prolly karma for something.....


Joe.

> wrote in message

 

Sorry not exactly right. It was just a bad guess at current drain. Both usb
cables ortho are the same make have different guage cable for the power go
figure. The WG111v2 dies just below 4.4volts so i found.

You fergot to factor in the weathering of the connectors...

The local ham radio club noticed a spike in current drain when a can was
used. could you factor that in too?

J.